1 2012-06-06 00:07:47 <BlueMatt> rebroad: ping
 2 2012-06-06 00:14:29 <micah> hi, i'm trying to get debian's version of bitcoind up to the latest version, but it is having trouble running some tests and I'm looking for some help in debugging that
 3 2012-06-06 00:14:54 <micah> if anyone is familiar with this, here is the build log of the most recent attempt to build it on i386: https://buildd.debian.org/status/fetch.php?pkg=bitcoin&arch=i386&ver=0.6.2.2-1&stamp=1336915445
 4 2012-06-06 00:23:29 <D34TH> micah missing some json files
 5 2012-06-06 00:26:09 <micah> D34TH: they seem to be there, but the path I see in the source is not where they are
 6 2012-06-06 00:27:23 <micah> Array tests = read_json("script_valid.json");
 7 2012-06-06 00:27:34 <micah> when it is actually in test/data/script_valid.json
 8 2012-06-06 00:27:55 <D34TH> edit source
 9 2012-06-06 00:27:56 <D34TH> :D
10 2012-06-06 00:28:55 <micah> D34TH: oddly, its succeeding on other architectures, with the same source, which I do not understand
11 2012-06-06 00:46:25 <luke-jr> micah: the path for the tests is hard-coded for now - is i386 moving it?
12 2012-06-06 00:46:41 <luke-jr> see also 9e71a5c
13 2012-06-06 00:47:20 <micah> luke-jr: i'm not sure... i just built it fine on an i386 machine, so that makes me wonder if there is a build machine problem
14 2012-06-06 00:49:35 <micah> luke-jr: is git://github.com/gavinandresen/bitcoin-git.git the 'canonical' upstream git repo?
15 2012-06-06 00:50:01 <luke-jr> micah: no, that's Gavin's development git repo
16 2012-06-06 00:50:47 <luke-jr> git://github.com/bitcoin/bitcoin.git is development/master and git://gitorious.org/+bitcoin-stable-developers/bitcoin/bitcoind-stable.git is stable branches
17 2012-06-06 00:56:32 <micah> luke-jr: looks like 9371a5c was in 6.1, i'm trying to build 6.2
18 2012-06-06 00:57:16 <luke-jr> micah: yes, I just mentioned that commit since it's touching all the relevant code to investigate :p
19 2012-06-06 05:06:57 <dk5> when a block is being mined and different nonces are tried (starting from 0), does the nonce restart each time the merkle root changes?
20 2012-06-06 05:08:19 <Joric_> also is it possible that we'll not find a valid hash after all? nonce is a short number
21 2012-06-06 05:09:35 <dk5> the documentation on the bitcoin site says something about nonce overflow and a field called extraNonce, but I haven't seen any information on this
22 2012-06-06 05:10:05 <dk5> https://en.bitcoin.it/wiki/Block_hashing_algorithm
23 2012-06-06 05:31:21 <gmaxwell> Joric_: yes, most of the time you don't find a solution for a single nonce span.
24 2012-06-06 05:32:15 <rebroad> BlueMatt: pong
25 2012-06-06 05:33:11 <dk5> gmaxwell: so what happens in that case?
26 2012-06-06 05:33:47 <gmaxwell> dk5: you getwork again, and either the timestamp has updated, or you'll get work with an incremented extranonce.
27 2012-06-06 05:35:39 <rebroad> anyone know what causes a "EXCEPTION: NSt8ios_base7failureE
28 2012-06-06 05:36:36 <dk5> gmaxwell: let's say I was a miner - isn't there an incentive not to add transactions to your block, since adding new transactions would change the merkle root?
29 2012-06-06 05:36:40 <gmaxwell> rebroad: it believe it's cause by the index referencing blocks past the end of the blk file, triggered by unclean shutdowns where the database update makes it to disk before the blocks.
30 2012-06-06 05:36:59 <gmaxwell> dk5: none at all. Did you see what I wrote to you in #bitcoin-mining.
31 2012-06-06 05:37:02 <gmaxwell> ?
32 2012-06-06 05:37:27 <gmaxwell> 00:30 < gmaxwell> dk5: mining can be seen as a random process like throwing dice where the odds of solution are constant, it's not a search of the sort where the more you  check the higher marginal odds as you exhaust a space.
33 2012-06-06 05:37:31 <gmaxwell> 00:32 < gmaxwell> dk5: the extranonce is in the coinbase transaction, so updating it changes the merkle root.
34 2012-06-06 05:38:13 <gmaxwell> changing the nonce is just like changing the merkleroot. Any change at all just gives you another equal shot at producing a solution.
35 2012-06-06 05:38:22 <rebroad> gmaxwell: oh. thanks. I've been getting a few of those in the last 24 hours...
36 2012-06-06 05:39:20 <gmaxwell> There is no fixed number of solutions. E.g. you could have a run of nonces in a row that were solutions .. or go several nonce spans ... so it's not like looking for a needle in a haystack and someone changing the haystack
37 2012-06-06 05:39:34 <gmaxwell> it's like throwing dice trying to get a 1 and someonce changes your dice.
38 2012-06-06 05:39:47 <rebroad> gmaxwell: not sure what you mean by referencing blocks past the end of the file... is there a way to code to avoid it causing bitcoind to exit?
39 2012-06-06 05:40:40 <gmaxwell> rebroad: if I'm correct about the cause and I think I am but there may be more than one cause. Checkblocks should just catch this and then remove the last (N) blocks until its not past the end anymore.
40 2012-06-06 05:42:36 <dk5> gmaxwell: thanks. I'm going to try to digest what you just said and make sense of it.
41 2012-06-06 05:45:40 <weex> Joric_: I created a github for my tweaked version of your addrgen https://github.com/weex/addrgen
42 2012-06-06 05:49:14 <rebroad> gmaxwell: it's just done it again.... :-s  I seem to be in an ideal position to help bug-fix this one
43 2012-06-06 05:50:37 <rebroad> ah.. I think it's due to something I've changed...
44 2012-06-06 06:01:18 <Joric_> weex, yeah, cool
45 2012-06-06 06:04:39 <Joric> weex, mind that KEY() class supposed to be the interchangeable wrapper i also have a pure python version of it
46 2012-06-06 06:05:29 <Joric> libssl version is just shorter and faster, maybe 100 times faster than pure python